面向闪存系统的LDPC纠错码技术研究

发布时间：2018-03-23 06:14

本文选题：闪存系统　切入点：纠错码　出处：《哈尔滨工业大学》2016年硕士论文　论文类型：学位论文

【摘要】：闪存(Flash memory)系统在使用过程中会不可避免地受到各种干扰而产生误码,因此通常需要引入纠错码(Error Correcting Code,ECC)来保证数据存储的可靠性。LDPC码(Low Density Parity-Check Code)是一种线性分组纠错码,因为其具有接近香农限的纠错性能在通信领域得到了广泛应用,但是LDPC码译码复杂度高,译码过程消耗时间长,限制了其在闪存系统中的应用。本文针对上述实际问题,研究了闪存系统的物理特性和LDPC码的基本理论,在此基础上提出了以较小代价获取较高精度软信息的方法,以提高LDPC译码器的性能,并将该优化方法从二进制LDPC码扩展至多进制LDPC码。本文的主要研究内容和取得的成果如下:1、LDPC码原理研究与参数设计。LDPC码的几个主要模块:矩阵构造算法,编码算法以及译码算法,都有多种不同的算法实现方式。闪存系统有其特殊的物理特性,对纠错码的各项参数都有限制,比如固定页宽度、超高码率等。本文在完成LDPC码的原理研究和MATLAB仿真的基础上,对已有算法进行调整,重新选择编码的各项参数,设计符合闪存系统实际应用的编码方案。2、提出了基于闪存错误特征的联合译码策略。LDPC码译码器的迭代效率受软信息精度影响,而以往闪存系统纠错方案中获取软信息的时间、空间代价过大。本文在对闪存系统驻留错误数据进行总结的过程中,发现MLC闪存的存储单元中驻留错误引起的位跳变存在一定的规律。根据这种规律,可以在原始错误率的基础上计算位错误率,提高软信息精度,加快LDPC码译码的迭代收敛速度,即减少译码时间消耗,同时增强纠错能力。进行了二进制LDPC码的对比实验,验证了该优化策略能够显著提高二进制LDPC码的时间效率和纠错能力。3、研究了面向闪存系统的多进制LDPC码。多进制LDPC码是二进制LDPC码在高阶有限域GF(q)的扩展,其已经被证明具有比二进制LDPC码更强的纠错能力,但多进制LDPC码的译码复杂度比二进制LDPC码更高。本文研究了多进制校验矩阵的构造方法和译码算法优化过程。在完成多进制LDPC码仿真的基础上,将联合译码方案应用于多进制LDPC码,经实验验证联合译码策略对多进制LDPC的迭代效率和纠错能力均有较大幅度的提升。
[Abstract]:Flash memory) system is inevitably affected by various kinds of interference, so it is usually necessary to introduce error Correcting code (Ecc) to ensure the reliability of data storage. Low Density Parity-Check code is a linear block error correction code. The error-correcting performance near Shannon limit has been widely used in the field of communication, but the complexity of decoding LDPC codes is high and the decoding process takes a long time, which limits its application in flash memory system. The physical characteristics of flash memory system and the basic theory of LDPC code are studied. On this basis, a method of obtaining high precision soft information at a lower cost is proposed to improve the performance of LDPC decoder. And the optimization method is extended from binary LDPC code to multiary LDPC code. The main contents and results of this paper are as follows: research on the principle of LDPC code and several main modules of parameter design: Matrix construction algorithm. There are many different algorithms for encoding and decoding. Flash memory systems have their special physical properties and limit the parameters of error-correcting codes, such as fixed page width, On the basis of the principle research of LDPC code and MATLAB simulation, this paper adjusts the existing algorithms and re-selects the parameters of the code. The coding scheme .2which accords with the practical application of flash memory system is designed. A joint decoding strategy based on flash error features. The iterative efficiency of LDPC decoder is affected by the precision of soft information. However, the time of obtaining soft information in the error-correcting scheme of flash memory system in the past is proposed. In the course of summarizing the error data of flash memory system, we find that the bit jump caused by resident error in the memory cell of MLC flash memory has a certain rule. The bit error rate can be calculated on the basis of the original error rate, the accuracy of soft information can be improved, and the iterative convergence rate of LDPC decoding can be accelerated, that is to say, the decoding time is reduced, and the error correction ability is enhanced at the same time. The contrast experiment of binary LDPC codes is carried out. It is verified that the optimization strategy can significantly improve the time efficiency and error-correcting ability of binary LDPC codes, and the multi-ary LDPC codes for flash memory systems are studied. The multi-ary LDPC codes are an extension of binary LDPC codes in high-order finite field. It has proven to be more error-correcting than binary LDPC codes, But the decoding complexity of the multiary LDPC code is higher than that of the binary LDPC code. This paper studies the construction method of the multiary check matrix and the optimization process of the decoding algorithm. On the basis of the simulation of the multiary LDPC code, The joint decoding scheme is applied to the multiary LDPC codes. The experimental results show that the joint decoding strategy can greatly improve the iterative efficiency and error-correcting ability of the multiary LDPC.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TN911.22

【相似文献】